Embed Size (px)
Citation preview
genesG C A T
T A C G
G C A T
Article
Trade in Zambian Edible Orchids—DNABarcoding Reveals the Use of UnexpectedOrchid Taxa for Chikanda
Sarina Veldman 1,* , Seol-Jong Kim 1 , Tinde R. van Andel 2 , Maria Bello Font 3,Ruth E. Bone 4, Benny Bytebier 5 , David Chuba 6, Barbara Gravendeel 2,7,8 ,Florent Martos 5,9 , Geophat Mpatwa 10, Grace Ngugi 5,11, Royd Vinya 10,Nicholas Wightman 12, Kazutoma Yokoya 4 and Hugo J. de Boer 1,2
1 Department of Organismal Biology, Systematic Biology, Uppsala University, Norbyvägen 18D,75236 Uppsala, Sweden; [email protected] (S.-J.K.); [email protected] (H.J.d.B.)
2 Naturalis Biodiversity Center, P.O. Box 9517, 2300 RA Leiden, The Netherlands;[email protected] (T.R.v.A.); [email protected] (B.G.)
3 Natural History Museum, University of Oslo, Postboks 1172, Blindern, 0318 Oslo, Norway;[email protected]
4 Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3AB, UK; [email protected] (R.E.B.);[email protected] (K.Y.)
5 Bews Herbarium, School of Life Sciences, University of KwaZulu-Natal, Pr. Bag X01, Scottsville 3209,South Africa; [email protected] (B.B.); [email protected] (F.M.); [email protected] (G.N.)
6 Department of Biological Sciences, University of Zambia, Box 32379 Lusaka, Zambia; [email protected] Institute of Biology Leiden, Leiden University, P.O. Box 9505, 2300 RA Leiden, The Netherlands8 University of Applied Sciences Leiden, Zernikedreef 11, 2333 CK Leiden, The Netherlands9 Institut de Systématique, Evolution, Biodiversité (ISYEB), Muséum national d’histoire naturelle, CNRS,
Sorbonne Université, EPHE, CP50, 45 rue Buffon 75005 Paris, France10 School of Natural Resources, The Copperbelt University, PO Box 21692 Kitwe, Zambia;
[email protected] (G.M.); [email protected] (R.V.)11 East African Herbarium, National Museums of Kenya, P.O. Box 40658-00100 Nairobi, Kenya12 Homegarden Landscape Consultants Ltd., P/Bag 30C, Chilanga, Lusaka, Zambia;
[email protected]* Correspondence: [email protected]; Tel.: +46-737828087
Received: 24 October 2018; Accepted: 22 November 2018; Published: 30 November 2018 �����������������
Abstract: In Zambia, wild edible terrestrial orchids are used to produce a local delicacy called chikanda,which has become increasingly popular throughout the country. Commercialization puts orchidpopulations in Zambia and neighbouring countries at risk of overharvesting. Hitherto, no studyhas documented which orchid species are traded on local markets, as orchid tubers are difficult toidentify morphologically. In this study, the core land-plant DNA barcoding markers rbcL and matKwere used in combination with nrITS to determine which species were sold in Zambian markets.Eighty-two interviews were conducted to determine harvesting areas, as well as possible sustainabilityconcerns. By using nrITS DNA barcoding, a total of 16 orchid species in six different genera couldbe identified. Both rbcL and matK proved suitable to identify the tubers up to the genus or familylevel. Disa robusta, Platycoryne crocea and Satyrium buchananii were identified most frequently andthree previously undocumented species were encountered on the market. Few orchid species arecurrently listed on the global International Union for the Conservation of Nature (IUCN) Red List.Local orchid populations and endemic species could be at risk of overharvesting due to the intensiveand indiscriminate harvesting of chikanda orchids, and we therefore encourage increased conservationassessment of terrestrial African orchids.
Genes 2018, 9, 595; doi:10.3390/genes9120595 www.mdpi.com/journal/genes
Genes 2018, 9, 595 2 of 15
Keywords: CITES; chikanda; conservation; DNA barcoding; orchids; species identification
1. Introduction
Terrestrial orchids have been used for medicinal and culinary purposes for centuries [1], with themost notable example being the use of orchid tubers to make salep, a traditional Turkish creamydrink or ice cream, consumed in Asia Minor and several countries on the Balkan peninsula [1–4].In south-eastern Africa, terrestrial orchid tubers are mixed with peanut flour, salt, baking soda andchili powder to make a traditional Zambian meat-like cake known as chikanda or African polony [5–7].Although initially not highly regarded [8], chikanda has more recently become popular throughout thecountry. It is sold as a snack along the streets, on markets, in supermarkets and on the menu of high-endrestaurants [9] and recipes; in addition, cooking tutorial videos can be found online [10]. Orchids usedfor chikanda are harvested exclusively from the wild, and although it is unlikely that traditional villageconsumption poses a serious threat to orchid populations, the increased popularity and subsequentcommercialization of chikanda has led to the exhaustion of Zambian orchid resources [7]. Collectingtubers means the end of a perennial and generally long-lived orchid, since the entire plant is removedin the harvesting process.
Soweto market in the Zambian capital Lusaka is the hub of the chikanda trade. Surveys performedon this market have shown that a large part of the chikanda tubers sold are sourced from Tanzania andthat Zambian chikanda orchids are collected from the Luwingu, Mporokoso and Kasama districts inthe Northern Province and Serenje in the Central province [6,7]. According to local chikanda vendors,another region with a flourishing chikanda trade is the Kitwe region in the Copperbelt Province, but sofar, no surveys have been performed there. Despite international legislation initiated by the Conventionon International Trade in Endangered Species of Wild Fauna and Flora (CITES) banning cross-bordertrade, an estimated 2–4 million orchid tubers are transported annually from Tanzania to Zambia [7,11].Import from the surrounding countries of Angola, Democratic Republic of the Congo (DRC), Malawiand Mozambique is also documented [7,9]. Orchid species originally reported as ingredients forchikanda are Disa robusta N.E.Br. and Satyrium buchananii Schltr. [11,12], whereas at least 32 speciesbelonging to the genera Brachycorythis, Disa, Eulophia, Habenaria, Roeperocharis and Satyrium wererecently suggested to be used for chikanda production based on collections in the field [11,13–18] andone metabarcoding study of ready-made chikanda cakes [18]. To date, however, no study has identifiedthe orchids traded at the local markets, since the tubers lack sufficient morphological characters fortaxonomic identification to species level [7,18]. Local classification systems categorize the tubers basedon texture, harvesting locality, soil color and phenology, but these are not likely to be congruent withscientific classifications [5,14].
Knowing which orchid species are currently being collected for the expanding chikanda tradeenables the identification of species susceptible to overharvesting and can inform conservationplanning. Molecular methods such as DNA barcoding can be applied to identify samples whenmorphological diagnostic characters for identification are lacking [19]. DNA barcoding andmetabarcoding has proven to be effective in the authentication of commercial wood species (Jiao, 2018),medicinal plants [20] and salep-producing orchids on Iranian markets [4]. The analysis of ingredientsin Tanzanian chikanda cake with DNA metabarcoding revealed the presence of 21 different orchidspecies [18], but a DNA barcoding approach has not yet been applied to individual orchid tubers usedto make this product. The aim of this study was to test to what extent the use of standard molecularmarkers yields robust identification of chikanda orchid tubers traded on Zambian markets. Molecularidentification can enable the mapping of the harvesting and trade of specific Zambian orchid speciesand facilitate the identification and implementation of targeted conservation strategies. Within thatframework, this study also aimed to identify conservation issues associated with the chikanda trade,and addresses the following questions: (1) Which species are used for chikanda production in the
Genes 2018, 9, 595 3 of 15
Lusaka and Kitwe districts of Zambia, and what is their geographic origin? (2) Can chikanda tubers beidentified up to species level using DNA barcoding? (3) How do local classification systems relate toscientific species concepts? (4) What are the main conservation issues associated with chikanda tradein the Lusaka and Kitwe districts?
2. Materials and Methods
2.1. Interviews and Sample Collection
Fieldwork in Zambia was conducted in 2016 in the Kitwe, Kalulushi, Luanshya, Ndola, Mufulira,Chingola and Chililabomwe districts of the Copperbelt Province, the Kapiri Mposhi and Serenjedistricts in the Central Province, and in the capital Lusaka (Figure 1). Semi-structured interviews wereconducted with harvesters, middlemen and vendors to obtain insight into chikanda commercialization,harvesting times, preferences and availability. The questionnaires consisted of three sections, one oninformant and interview characteristics, one with general questions about chikanda posed to allinformants, and a third section with questions more specifically designed for each interviewee category:harvester, middleman and vendor. All research was conducted in accordance with the InternationalSociety of Ethnobiology Code of Ethics [21]. Ethical clearance was obtained from the Humanitiesand Social Science Research Ethics Committee of the University of Zambia. The interviews wereperformed in English or Bemba, with a translator affiliated with the University of Zambia and theCopperbelt University. Informants were selected using the snowball technique [22] by asking peoplewhether they could direct us to people harvesting or selling chikanda. All informants were providedwith information about the study and signed a prior informed consent sheet. Fieldwork took placeduring June and July, the peak season for chikanda [7], to ensure the collection of both fresh anddried chikanda tubers on the market and in the field. A collection was made each time a specificvernacular chikanda type was bought from a specific vendor, and assigned a collection number (SJK1,SJK2, etc.). Each individual tuber within the collection received a subsample number within thatcollection (SJK1.1, SJK1.2, SJK 1.16, etc.). The fresh tubers were sliced and stored with silica gel inplastic bags. All chikanda samples were brought to Sweden under a CITES inter-institutional exchangeagreement between the University of Zambia (ZM001) and the Botany Section of the EvolutionaryMuseum in the Evolutionary Biology Center in Uppsala (SE009). Export permission was obtainedfrom the Zambian CBD and Nagoya Protocol focal point at the Ministry of Natural Resources andEnvironmental Protection. Upon arrival in Sweden, some of the chikanda tubers had sprouted.Those were transferred to the Uppsala Botanical Garden for cultivation and subsequent sampling offresh leaf tissue for DNA barcoding as well as morphological identification.
Genes 2018, 9, x FOR PEER REVIEW 3 of 18
2.1. Interviews and Sample Collection
Fieldwork in Zambia was conducted in 2016 in the Kitwe, Kalulushi, Luanshya, Ndola, Mufulira, Chingola and Chililabomwe districts of the Copperbelt Province, the Kapiri Mposhi and Serenje districts in the Central Province, and in the capital Lusaka (Figure 1). Semi-structured interviews were conducted with harvesters, middlemen and vendors to obtain insight into chikanda commercialization, harvesting times, preferences and availability. The questionnaires consisted of three sections, one on informant and interview characteristics, one with general questions about chikanda posed to all informants, and a third section with questions more specifically designed for each interviewee category: harvester, middleman and vendor. All research was conducted in accordance with the International Society of Ethnobiology Code of Ethics [21]. Ethical clearance was obtained from the Humanities and Social Science Research Ethics Committee of the University of Zambia. The interviews were performed in English or Bemba, with a translator affiliated with the University of Zambia and the Copperbelt University. Informants were selected using the snowball technique [22] by asking people whether they could direct us to people harvesting or selling chikanda. All informants were provided with information about the study and signed a prior informed consent sheet. Fieldwork took place during June and July, the peak season for chikanda [7], to ensure the collection of both fresh and dried chikanda tubers on the market and in the field. A collection was made each time a specific vernacular chikanda type was bought from a specific vendor, and assigned a collection number (SJK1, SJK2, etc.). Each individual tuber within the collection received a subsample number within that collection (SJK1.1, SJK1.2, SJK 1.16, etc.). The fresh tubers were sliced and stored with silica gel in plastic bags. All chikanda samples were brought to Sweden under a CITES inter-institutional exchange agreement between the University of Zambia (ZM001) and the Botany Section of the Evolutionary Museum in the Evolutionary Biology Center in Uppsala (SE009). Export permission was obtained from the Zambian CBD and Nagoya Protocol focal point at the Ministry of Natural Resources and Environmental Protection. Upon arrival in Sweden, some of the chikanda tubers had sprouted. Those were transferred to the Uppsala Botanical Garden for cultivation and subsequent sampling of fresh leaf tissue for DNA barcoding as well as morphological identification.
Figure 1. Map of Zambia with an overview of interview localities and reported provenance of the chikanda tubers. Dot size corresponds to the number of informants.
Commented [MX1]: incorrect ref order, 22 detected after 20.
You jumped the numbers in between.
Figure 1. Map of Zambia with an overview of interview localities and reported provenance of thechikanda tubers. Dot size corresponds to the number of informants.
Genes 2018, 9, 595 4 of 15
2.2. Reference Taxon Sampling
Herbarium specimens were collected with associated silica-dried material for DNA extractionand spirit collections during fieldwork in Zambia in January and February 2017. All material wascollected and exchanged in accordance with national and international legislation. The collectionswere deposited at the Division of Forest Research (Kitwe, Zambia) and RBG Kew (UK) and fieldidentifications verified at the Bews Herbarium (South Africa). A total of 94 novel Orchidaceae referencevouchers were collected for this study, representing 4 Brachycorythis, 9 Disa, 16 Habenaria, 6 Satyriumspecies and 26 species of other orchid genera. Voucher specimens of all taxa sampled are listed inTable S1 (Supplementary Materials). In addition, 88 nrITS, 71 matK and 45 rbcL Habenaria sequencesgenerated for a forthcoming phylogenetic study [21], and 510 nrITS, 522 matK and 213 rbcL sequencescorresponding to 311, 325 and 100 taxa in the previously mentioned orchid genera downloaded fromGenBank were included in the reference database.
2.3. From Sample to Sequence
Out of the 1284 individual tubers in 48 different sample collections, 304 samples were selected forDNA extraction. To assess the orchid species present in each collection, a few tubers per sample (2–8)were extracted if the sample was morphologically homogenous, whereas more (6–33) were selectedif the sample seemed diverse. In the selection process, we aimed to select tubers of different shapesand sizes to cover the potential species diversity in each collection. Since the selection appearedto be representative for each collection, it was not deemed necessary to extract DNA from eachindividual tuber in all market collections. DNA was extracted using a CTAB protocol [23] modifiedwith 3 to 5 extra washing steps with STE buffer (0.25 M sucrose, 0.03 M Tris, 0.05 M EDTA) [4,24]to reduce the gelatinization effect of the large amount of polysaccharides in the starch-rich orchidtubers. Total DNA was stored in 70–100 µL 10 mM Tris-HCl buffer, pH 8.0. DNA concentration wasmeasured with a Qubit 3.0 fluorometer (Thermo Fisher Scientific, Oakwood, GA, USA). The core landplant barcoding markers rbcL and matK were amplified using the primers and protocols described inKress et al. [25] and Dunning and Savolainen [23] respectively. The reactions were performed in a totalreaction volume of 25 µL with 14.725 µL ddH2O, 2.5 µL DreamTaq Buffer (Thermo Fisher Scientific,Oakwood, GA, USA), 0.5 µL 25 mM dNTP, 0.65 µL 2% bovine serum album (BSA), 0.125 µL DreamTaqPolymerase, 2.5 µL 5 pmol forward and reverse primer and 1.5 µL template DNA. Nuclear ribosomalITS was amplified using the Sun et al. [26] primers and protocol in a total reaction volume of 25 µLcontaining 15.25 µL ddH2O, 2.5 µL DreamTaq Buffer (Thermo Fisher Scientific), 0.5 µL 25 mM dNTP,0.125 µL 2% BSA, 0.125 µL DreamTaq Polymerase (Thermo Fisher Scientific), 2.5 µL 5 pmol forwardand reverse primer and 1.5 µL template DNA. For ITS, an additional protocol was used with Q5high-fidelity polymerase; reactions were performed in a total reaction volume of 23.5 µL including10.875 µL ddH2O, 5 µL Q5 reaction buffer, 0.5 µL 25 mM dNTP, 5 µL Q5 GC enhancer, 0.125 µLQ5 high-fidelity polymerase, 1.5 µL 5 pmol forward and reverse primer and 0.5 µL template DNA.The PCR program for the ITS primers in combination with the Q5 polymerase was an initial heatingstep of 30s at 98 ◦C, 35 cycles of 10s 98 ◦C, 30s 56 ◦C, 30s 72 ◦C, and a final elongation of 2min at 72 ◦C.PCR products were cleaned using eight-times diluted ExoSAP (Thermo Fisher Scientific) and analysedon an ABI3730XL Sanger sequencer by Macrogen Europe (Amsterdam, The Netherlands). The obtainedtrace files were assembled using Pregap4 and Gap4 [27] as implemented in the Staden package [28].Sequences shorter than 200 bp were discarded from the analysis and all sequences were depositedin NCBI GenBank (Table S2, Supplementary Materials). The NCBI BLAST algorithm was used toassess the identification of all the obtained query sequences using the Python BLAST tabular parserscript (https://github.com/SLAment/Genomics/blob/master/BLAST/BLAST_tabularparser.py).Similarity scores, query coverage, expect value (E-value), and maximum identity percentage werecalculated, and all the information of the top 5 hits were automatically mined and tabulated per marker(Tables S3–S5, Supplementary Materials) and summarized per individual tuber sample (Table S6,Supplementary Materials).
Genes 2018, 9, 595 5 of 15
2.4. Phylogenetic Analysis and Species Identification
Alignments for nrITS, matK and rbcL were made using AliView [29], combing the query sequencesand local reference databases consisting of sequences from NCBI GenBank and reference collectionsfrom fieldwork and unpublished data from collaborators [21]. Species identification was performedusing the Bayesian implementation of the Poisson tree processes model (bPTP), as it has been shownto outperform the generalized mixed Yule coalescent (GMYC) approach as well as OperationalTaxonomic Unit (OTU)-picking methods when evolutionary distances between species are shortenough [30]. For all alignments, a maximum likelihood (ML) search for the best-scoring tree wasperformed using the RAxML web server [31] to generate input trees for species identification analysisusing bPTP [30]. The default setting, GTRCATI, was used to implement the CAT approximation,and the final tree was evaluated using the traditional GTR model. The bPTP.py script setting was100,000 Markov chain Monte Carlo (MCMC) iterations for all trees, a sampling interval thinning valueof 100, a burn-in of 25%, and a random seed of 1234. No outgroup was defined. The generatedconvergence curve was visually confirmed. The description of each voucher is built up as follows:‘sample #_BLAST search result_identification %_reported city/region of the origin_the country,’;for example: ‘SJK04.09_S.buchananii_97.893_Mwinilunga_Zambia’ (Figures S1–S3, SupplementaryMaterials). For the reference sequences, the accession number and species names are described forDisa and Satyrium. The reference Habenaria species and others are described as their species name andvoucher (sample) number.
3. Results
3.1. Market Surveys and Interviews
We visited 25 markets in ten Zambian cities: seven cities in the Copperbelt Province, two inthe Central Province, and one in Lusaka (Figure 1). Eighty-two persons involved in chikanda tradewere interviewed, of which 8 were harvesters, 44 middlemen, 29 vendors and one informant wasboth harvester and middleman. The term ‘harvesters’ refers to people collecting chikanda in the field,‘middlemen’ to people selling either dried or fresh chikanda orchid tubers on the market, and ‘vendors’to informal street vendors selling ready-made chikanda cakes contained in a basket carried on thehead. Since one of the harvesters also acted as middleman, in total, 83 individual interviews wereconducted: 72 participants were female and 10 were male. The ages of the respondents varied between18 and 78, with an average of 41 years. The majority of the respondents belonged to the Bemba tribe(62%), while the other respondents (38%) belonged to smaller ethnic groups. Most interviews wereconducted in the Copperbelt Province (55), eight in the Central Province and 19 in Lusaka.
3.2. Local Classification System
Fifty-three different vernacular names for the various chikanda tuber types were recorded duringthe interviews. The most common way to distinguish between chikanda tubers was by using theterms original (myala) and fake (mbwelenge or msekelele). In some cases, it was the shape of the tuberthat was used to differentiate between the different tubers: mshilamshila means root-like in Bembaand referred to the elongated, root-shaped tubers, whereas mampanda referred to the heart-shapedtubers. It also appeared common to use the origin of the tuber as a trade name: mwinilunga, chozi,luwingu and kasama are, for example, all Zambian city names, sumbawanga and iringe refer to Tanzaniancities (Sumbawanga and Iringa) and angola refers to one of the countries bordering Zambia. In somecases, the chikanda tubers were sold pre-mixed, whereas other vendors marketed the different types oftubers separately. The morphology and size of the tubers varied both within and between collectionsof a certain chikanda type. Tubers could be heart-shaped, rounded, egg-shaped to elongated andalmost root-like. The largest tubers were the heart-shaped ones, which could be up to 5 cm long and6 cm wide. The elongated tubers were up to 9 cm long with a maximum width of 2 cm. Harvesters,middlemen and vendors themselves indicated that they distinguish the tubers based on the size of
Genes 2018, 9, 595 6 of 15
the granules inside the tubers, which can be large or small, and in some tubers, a concentric ring wassaid to be present. Figure 2 illustrates the different tuber types and ready-made chikanda that wereencountered on the local markets, as well as some of the orchid species producing these tubers.
Genes 2018, 9, x FOR PEER REVIEW 6 of 18
based on the size of the granules inside the tubers, which can be large or small, and in some tubers, a concentric ring was said to be present. Figure 2 illustrates the different tuber types and ready-made chikanda that were encountered on the local markets, as well as some of the orchid species producing these tubers.
Figure 2. Chikanda tubers, cake and orchids. (a) Myala—real chikanda; (b) Mbwelenge—fake chikanda; (c) Mshilamshila—supposedly Brachycorythis sp.; (d) Mampanda. (e–g); Chikanda cake; (h) Disa robusta; (i) Disa welwitschii; (j) Platycoryne crocea; (k) Satyrium carsonii; (l) Satyrium buchananii; (m) Satyrium kitimboense; (n) Brachycorythis cf. friesii; Photographs (a–g) by Seol-Jong Kim, (h) by Robert v.
Figure 2. Chikanda tubers, cake and orchids. (a) Myala—real chikanda; (b) Mbwelenge—fake chikanda;(c) Mshilamshila—supposedly Brachycorythis sp.; (d) Mampanda. (e–g); Chikanda cake; (h) Disarobusta; (i) Disa welwitschii; (j) Platycoryne crocea; (k) Satyrium carsonii; (l) Satyrium buchananii; (m)Satyrium kitimboense; (n) Brachycorythis cf. friesii; Photographs (a–g) by Seol-Jong Kim, (h) by Robert v.Blittersdorff, (i,k,l) by Nicholas Wightman, (j) by Warren McCleland, (m) by Ruth E. Bone and (n) bySarina Veldman.
3.3. Chikanda Trade and Availability
Many participants had a relatively long experience (an average of 13.5 years) in the chikanda trade.They indicated that they were asked by family, friends or neighbors to get involved in the chikandabusiness, or simply because it seemed to be a profitable industry. Some of the participants, especiallyharvesters, stated that they started collecting chikanda because the plants were easily accessible andwere found growing close to their areas of residence. In general, people involved in the chikandatrade indicated that this work alone did not suffice to fully support themselves and their families,and these traders therefore supplemented their income by trading additional natural products such
Genes 2018, 9, 595 7 of 15
as fruits, maize, groundnuts, beans, mushrooms, snacks, herbs, and kapenta (a dried fish from LakeTanganyika). In Lusaka, however, income generated with chikanda trade seemed to be sufficient forsubsistence, and this suggests that there are significant local differences in profits generated in thechikanda business.
The chikanda trade was structured in several different ways. Some of the middlemen loadedthe orchid tubers in trucks directly from the harvesting areas and brought them to larger cities suchas Lusaka and Kitwe, whereas others relied on agents to gather a specified amount of chikandatubers, which they paid for through the East-African mobile payment system M-Pesa and receivedas cargo from one of the local buses. In addition to the Tunduma and Nakonde markets that arethe trade hubs between Tanzania and Zambia, other centers of trade were identified on the borderwith the DRC (Chililabombwe and Kasumbalesa) and Angola (Mwinilunga), where both chikandatubers and ready-made chikanda cake were sold. An overview of all interview localities and thereported provenance of the chikanda tubers is given in Figure 1. Most of the participants indicated thatchikanda plants are becoming locally rare. Middlemen emphasized the decrease of quantity, whereasharvesters were concerned about the decline in both quantity and quality (size and preferred chikandatype), which may also depend on the species. The participants from urban areas stated that access tochikanda tubers was managed by the chief of each tribe, who seasonally designated the harvestabledambo (wet meadow) within the chiefdom so that the collectors could maintain the quality of theharvests. Nevertheless, interviewed harvesters found in the dambo areas of Serenje (Central Province)claimed to be free to harvest tubers whenever available.
3.4. Molecular Identification of Traded Orchids
During the DNA extraction, many samples formed a thick jelly-like layer in the extraction tubes,despite repeated washing steps with STE buffer. This resulted in a very small water phase and islikely to have negatively influenced the downstream steps of the extraction process. The average DNAconcentration of the chikanda tuber samples was 4.96 ng/µL, while the average DNA concentrationfrom leaf samples from chikanda orchids was 28.5 ng/µL. Out of the 304 samples selected for DNAextraction, 232 samples produced detectable DNA. Amplification was attempted for each barcodingmarker for all of the samples. A nrITS sequence was obtained for 159 samples, rbcL for 117 samples andmatK only for 45 samples. Sequences from all three markers were obtained for 40 samples, sequencesfrom two markers for 58 samples, and 55 samples only yielded sequences for a single marker. Analysisof the inter- and intraspecific variation of nrITS was performed for 141 sequences representing 124 Disaspecies, 73 sequences representing 59 Habenaria species and 110 sequences representing 67 Satyriumspecies. In the case of matK 135 sequences belonging to 122 Disa species were included, 507 Habenariasequences belonging to 239 species and 116 sequences belonging to 60 Satyrium species. For rbcL,the available reference material was quite limited: 45 sequences for 37 Habenaria species, eight Satyriumsequences for two species, and four Disa sequences for four species.
A graphical overview of the inter- and intraspecific variation per genus and per marker can befound in Figure 3. The interspecific variation for nrITS was significantly higher as compared withmatK and rbcL in all three genera: on average, 10.2% in Disa, 9.51% in Satyrium and 8.75% in Habenaria.For matK, the interspecific variation was 2.39% for Disa, 2.79% for Habenaria and 1.39% for Satyrium.The intraspecific distances for nrITS and matK were, respectively, 1.37% and 0.61% for Disa, 1.98%and 0.36% for Habenaria and 1.45% and 0.48% for Satyrium, The limited reference sequences that wereavailable for rbcL showed little pairwise interspecific distances (1.44% for Disa, 0.82% for Habenariaand 1.13% for Satyrium) and only allowed for the intraspecifc distance calculation of Habenaria (0.38%),indicating that rbcL is not suitable as a barcode for the species-level identification of chikanda orchids.The calculated thresholds were subsequently used to evaluate the identifications made with the onlineand standalone Basic Local Alignment Search Tool (BLAST).
Genes 2018, 9, 595 8 of 15
Genes 2018, 9, x FOR PEER REVIEW 8 of 18
For matK, the interspecific variation was 2.39% for Disa, 2.79% for Habenaria and 1.39% for Satyrium. The intraspecific distances for nrITS and matK were, respectively, 1.37% and 0.61% for Disa, 1.98% and 0.36% for Habenaria and 1.45% and 0.48% for Satyrium, The limited reference sequences that were available for rbcL showed little pairwise interspecific distances (1.44% for Disa, 0.82% for Habenaria and 1.13% for Satyrium) and only allowed for the intraspecifc distance calculation of Habenaria (0.38%), indicating that rbcL is not suitable as a barcode for the species-level identification of chikanda orchids. The calculated thresholds were subsequently used to evaluate the identifications made with the online and standalone Basic Local Alignment Search Tool (BLAST).
Figure 3. Boxplots showing the inter- and intraspecific variation for Disa (a), Habenaria (b) and Satyrium (c) based on genetic diversity.
In total, 15 orchid species were identified using DNA barcoding: Brachycorythis sp. SJK7, Disa caffra Bolus, D. celata Summerh., D. robusta N.E.Br., D. satyriopsis Kraenzl., D. welwitschii Rchb. f., Disa sp. SJK4.1., Habenaria sp. SJK31.15, H. aff. helicoplectrum Summerh., Habenaria cf. sp. DO112, Platycoryne crocea Rolfe, Satyrium buchananii Schltr., S. carsonii Rolfe, and S. kitimboense Kraenzl., as well as one species in an unidentified genus, which seems to be closely related to Habenaria based on
Figure 3. Boxplots showing the inter- and intraspecific variation for Disa (a), Habenaria (b) and Satyrium(c) based on genetic diversity.
In total, 15 orchid species were identified using DNA barcoding: Brachycorythis sp. SJK7, Disacaffra Bolus, D. celata Summerh., D. robusta N.E.Br., D. satyriopsis Kraenzl., D. welwitschii Rchb. f.,Disa sp. SJK4.1., Habenaria sp. SJK31.15, H. aff. helicoplectrum Summerh., Habenaria cf. sp. DO112,Platycoryne crocea Rolfe, Satyrium buchananii Schltr., S. carsonii Rolfe, and S. kitimboense Kraenzl., as wellas one species in an unidentified genus, which seems to be closely related to Habenaria based on thesimilarity-based BLAST identification. Additionally, one orchid that flowered from a sprouting tuberwas identified based on morphology as Brachycorythis cf. friesii (Schltr.) Summerh.
bPTP analysis for matK and rbcL showed a lumping of several supposedly different specieswithin one clade on several occasions (Figures S1 and S2, Supplementary Materials). In the caseof nrITS, the bPTP outcome tree often reflected the expected species boundaries, although theposterior probabilities on the nodes were often too low to determine with confidence if speciesidentification had been performed correctly (Figure S3, Supplementary Materials). Something thatcould be observed from both the nrITS BLAST identification, as well as the bPTP analysis, was thatsome of the identifications showed ambiguity: in collection SJK41, SJK44 and SJK46, some sampleswere identified unambiguously as Platycoryne crocea, whereas others showed a mix of Habenaria andPlatycoryne top hits. The matK BLAST identification showed an unambiguous P. crocea identificationfor these samples, despite the presence of several Habenaria species in the reference database anddespite the fact that matK shows a lower level of interspecific variation. A similar observation couldbe made for several Disa samples, but here a geographic pattern could be observed. The samplesthat could unambiguously be identified as D. robusta were all collected from Tanzania, the originalsource of the reference sequence. Some of the Zambian samples also showed D. robusta as the closestrelative, but not with a high enough percentage identity match to confirm this identification. The
Genes 2018, 9, 595 9 of 15
bPTP result in this case shows a lumped and mixed clade with several Disa species and several othersamples included, which confirms that sequence divergence is too limited to resolve the relationship.These samples were identified as Disa sp. 1, to reflect the fact that they all grouped together inthe same clade. In other cases, such as for certain Habenaria species, we named the closely relatedsequences in the identification, whereas in this case the clade was too large to allow for this, since itcontained the following sequences: Disa engleriana Kraenzl., D. erubescens Rendle., D. miniata Summerh.,D. ochrostachya Rchb.f., D. satyriopsis, D. ukingensis Schltr., D. verdickii De Wild., D. welwitschii, D. zombicaN.E.Br., an unidentified Disa species and several samples that showed a highest percentage identitymatch with D. robusta, D. satyriopsis and D. welwitschii. The posterior probability for this clade is only0.25, and combined with the posterior probabilities present on the within-species nodes, indicates a lotof uncertainty for this identification.
Overall, the most frequently encountered species were Satyrium buchananii, with 41 samples ineleven different collections, Platycoryne crocea, with 19 samples in three different collections, and Disarobusta, with 4–16 samples in two to seven different collections. Myala or original chikanda seems tocontain D. robusta, D. welwitschii and S. buchananii. Mbwelenge or fake chikanda seems to correspondonly to S. buchananii and mshilamshila to one or several Brachycorythis species. Kasebelele and kapapareferred to Habenaria spp., P. crocea, S. carsonii and S. kitimboense. The mixed collections contained all ofthe above-mentioned species as well as an unidentified Habenaria species. An overview of the localchikanda classification types, their collection numbers, identifications and number of samples can befound in Table 1.
Table 1. Overview of the different local chikanda classification types, their collections and the identifiedscientific species.
Vernacular Name Collections Reported Origin Barcoding IDs # Samples
Fungulwe SJK16 unknown Disa robusta 1
Iringe SJK17 TanzaniaSatyrium buchananii 1
Satyrium carsonii 1John White SJK39 Mporokoso, Zambia Satyrium buchananii 1
Kabula seke SJK46 Serenje, ZambiaHabenaria sp. (Clade H. schimperiana, H. kyimbilae,
H. microsaccos 1
Platycoryne crocea 7Kapapa SJK44 Mporokoso, Zambia Platycoryne crocea 6
Kasebelela, John White andMyala SJK41 Chinsali and Mporokoso, Zambia
and Tanzania
Habenaria cf sp. DO122 (Clade H. schimperiana,H. kyimbilae, H. microsaccos 4
Platycoryne crocea 4Platycoryne sp./Habenaria sp. 2
Kasebulela and Kapapa SJK11 Luwingu, Zambia Satyrium kitimboense 6Satyrium carsonii 5
Mbwelenge SJK5 Luwingu, Zambia Satyrium buchananii 11Satyrium sp. 1
SJK32 Serenje, Zambia Satyrium buchananii 6
MshilamshilaSJK7 Luwingu, Zambia Brachycorythis sp. 1
SJK12 Kawamba, ZambiaBrachycorythis sp. 1
Brachycorythis cf. friesii 1
MyalaSJK4 Mwinilunga, Zambia;
Disa robusta 4Disa welwitschii 1
Satyrium buchananii 4
SJK18 Sumbawanga, Tanzania Disa robusta 4Satyrium buchananii 1
Myala SJK37 Kawambwa, Zambia Satyrium buchananii 1
Myala and nampanda SJK21 Luapula, Zambia Disa welwitschii 2Satyrium buchananii 1
Ntonkonshi SJK25 Democratic Republic of Congo Disa robusta 1Sumbawanga SJK20 Sumbawanga, Tanzania Disa satyriopsis 1
Mixed SJK31 Serenje, Zambia
Disa caffra 1Disa robusta 1
Habenaria cf sp. DO122 (Clade H. schimperiana,H. kyimbilae, H. microsaccos 1
Satyrium buchananii 6
Unknown-mixed SJK19 Luwingu, Zambia Satyrium carsonii 1Habenaria aff. helicoplectrum (BB3151) 1
Unknown
SJK8 Mwinilunga, Zambia
Disa miniata 1Disa robusta 2
Disa welwitschii 1Satyrium buchananii 2
SJK9 Luwingu, Zambia Satyrium carsonii 1
SJK13 Kawamba, ZambiaDisa celata 1
Disa welwitschii 1Satyrium buchananii 3
Genes 2018, 9, 595 10 of 15
4. Discussion
4.1. Species Used for Chikanda
Using DNA barcoding as an identification tool for chikanda tubers sold on local Zambianmarkets has allowed us to determine for the first time which orchid species are sold on localmarkets. Previous studies identifying orchids used for chikanda relied on voucher collections madewith collectors in the field and their morphological identification, which requires a qualified orchidtaxonomist [6,11,13,16,17]. Additionally, relying on local harvesters for details on chikanda collectionmight not always lead to collection in the areas where actual intensive harvesting is taking place,since some harvesters do not like to divulge where the best places to harvest are, and in some initialharvesting areas, such as the Kitulo Plateau in Tanzania’s Southern Highlands, collection is nowprohibited [32]. The current study identified 16 orchid species present on the markets, includingat least three previously undocumented ones: Brachycorythis cf. friesii, Platycoryne crocea and anunidentified species in a genus, which appears to be closely related to Habenaria, but is not present inour reference database.
Orchids used for chikanda seem to be harvested from several provinces in Zambia, as well asat least two regions in Tanzania. Moreover, three international chikanda trade hubs in towns onor close to the border with the DRC (Chililabombwe and Kasumbalesa) and Angola (Mwinilunga)were identified, in addition to the already known trade hub Tunduma-Nakonde on the Tanzanianborder [7,11]. However, unlike what has been reported in other studies, no harvest from Malawi wasmentioned by the people interviewed in this study, which could mean that trade from this countryis currently not taking place. Another explanation is that this information is lost en route and thatonly Tanzania, being geographically closer and thus more easily accessible for the Bemba people,is mentioned as a region of origin for chikanda traded in Zambia.
4.2. DNA Barcoding Performance
Since the term DNA barcoding was coined in 2003 [19], a plethora of studies applying DNA(meta)barcoding has been performed ranging from retrieving orchids from paleoenvironments [33],preserved in mammoth dung [34], to the identification of Iranian orchid tubers used for salep [4]. In thisstudy, a combined use of the core plant markers matK and rbcL and the nuclear ribosomal ITS regionwas used to attempt the species-level identification of tubers traded on Zambian markets. Althoughgenetic distance calculations showed limited interspecific distances between closely related speciesfor all three barcoding markers, DNA barcoding allowed for species-level identification for severalof the frequently sold chikanda species. The data shows that the core land-plant DNA barcodingmarkers rbcL and matK were not suitable because of the limited variability between species (matK andrbcL), amplification problems (matK), and/or a limited sequence reference database (rbcL). Similarperformance has been reported in other barcoding studies and indicates that the use of these coreland plant markers might not be suitable for the analysis of samples with degraded DNA and for theidentification of closely related species [35–37]. nrITS was shown to be more suitable as a barcodemarker to distinguish between different chikanda species, but is not discriminative enough to enablereliable species level identification in certain orchid clades, such as the clade with Platycoryne crocea,P. buchananiana and Habenaria buchananii; the clade with H. schimperiana, H. kyimbilae and H. microsaccosand the clade containing Disa sp. 1. Another drawback of nrITS is the presence of multiple ITSparalogs in the ribosomal genome. Usually, these copies would show high similarities due to concertedevolution [38,39], but this is not always the case [40–43]. In our case, potentially different nrITSribotypes became fixed in different orchid populations, and having only one of them in our referencedatabase could lead to the unresolved identifications observed. Our bPTP results show that evennrITS has too low a resolution to reliably identify species with high posterior probability supportusing this method. It does demonstrate, however, how valuable the use of tree-based methods canbe, since it shows the relations between the sequences and can be used to determine if some of the
Genes 2018, 9, 595 11 of 15
unidentified samples are likely to belong to the same species or a different species within the samegenus. Even if no species-level identification can be made for these samples, it is possible to usethe clustering to determine the diversity of species used. Although several samples can only bereliably identified as Habenaria sp., we find that they are likely to belong to at least three differentspecies (Habenaria aff. helicoplectrum, Habenaria cf. sp. DO122 in the clade with H. schimperiana,H. kyimbilae and H. microsaccos and lastly the Platycoryne sp./Habenaria sp., which group togetherwith H. buchananii, P. buchananiana and P. crocea. Expansion of the reference database, by includingat least one individual per species, and preferably multiple individuals per species from differentpopulations and countries, could ultimately solve the remaining challenges, and this seems the wayforward for the identification of the traded chikanda tubers, as well as other species unidentifiableby morphology. Similar studies using DNA barcoding for the identification of unknown samplesshow comparable results for the employed barcoding markers. In a study on the identification oforchids used for salep, nrITS showed a sequencing success more than three times higher than matKas well as a two-fold higher species-level identification success [4]. Moreover, the similarity-basedapproach seemed to outperform the tree-based identification method (ML) in this study as well with57% and 39% species-level identifications, respectively. nrITS also shows the highest identificationperformance in studies on the identification of medicinal plants [44,45], which supports the idea thatit is advisable to add a more discriminative marker to the two core land-plant barcodes in studieswhere it is needed to distinguish between closely-related species. Moreover, our results stress the needfor a phylogenetically underpinned taxonomic framework, which is currently available for Disa [46]and Satyrium [47], but not yet for Habenaria and related genera. A remaining limitation is the limitedDNA extraction and amplification result, which could potentially be improved by experimenting withdifferent extraction methods and primer pairs that would allow the amplification of segments of thematK marker [48–50].
4.3. Local Versus Scientific Classification of Chikanda
The results of our identifications of chikanda orchids traded in Zambia show that the localclassification systems for chikanda are not congruent with the botanical classification of orchidspecies. The orchids sold on the markets were grouped according to area of origin, tuber consistencypreference, or shape of the tuber, but often the tubers offered for sale were mixtures. The differentlocal types of chikanda sometimes show a variation in orchid species that are identified within theselocal groupings. When we look at the chikanda type known as kapapa, for example, SJK44 containsPlatycoryne crocea, whereas SJK11, which is supposed to be a mix of kasebelela and kapapa, only seems tocontain Satyrium species. In the case of other chikanda types, there seems to be more consistency: myalaor real chikanda referred to Disa robusta, D. welwitschii and S. buchananii, mshilamshila samples wereidentified as Brachycorythis species and mbwelenge or fake chikanda was made of S. buchananii. However,our previous study on the analysis of Tanzanian chikanda cakes showed that the cake made with fakechikanda tubers also contained Disa miniata, Satyrium anomalum, S. comptum, S. elongatum, S. riparium, S.shirense and S. volkensii, indicating that there might be some differences between fake chikanda samplesas well [18]. It is well known in the literature that local species concepts are not necessarily congruentwith scientific classifications and that species might be subject to over- or under-differentiation [51,52].In this case, the grouping of the orchids according to the area of origin, shape or consistency preferenceor plainly under the general term chikanda is clearly a case of under-differentiation, as a much higherdiversity was retrieved when employing DNA barcoding. In order to more reliably identify the orchidspecies used for a particular chikanda type, more samples per local classification need to be analyzed.
4.4. Orchid Availability and Conservation
In recent decades, chikanda has made a remarkable leap in popularity. The first record of chikandause made by Audrey Richards [8] described the relish as a poor man’s food, eaten in times of famine.Recent studies, from 2002 onwards, show that chikanda has emerged as a Zambian snack which
Genes 2018, 9, 595 12 of 15
is popular throughout the country. Studies on chikanda report that, with this rise in popularity,the orchid harvest has escalated and is pressuring local Zambian orchid populations as well as those inneighboring countries [7,9,11,53]. Many people involved in the chikanda trade indicated that chikandaplants were becoming scarce, and many were concerned about both the quality as well as the quantity ofthe orchids available. Our study also confirmed a significant international trade network for chikandasources in several regions in Tanzania, as well as in the Democratic Republic of Congo and Angola.In our current study, we found at least 16 different orchid species sold as chikanda on the Zambianmarkets and an overview of previous studies contains 46 species reported to be used for chikanda(Table S7). This brings us from the use of an initial two orchid species reported for chikanda [12] to atotal of 49 species in eight different genera. The increased harvesting pressure, in combination withthe indiscriminate harvesting and use of many more species than earlier assumed, poses a threat tonearly half of the terrestrial orchids occurring in these regions. Despite the establishment of KituloNational Park in Tanzania, with orchid conservation as a prime concern, it seems that harvestingcontinues even there, since iringe tubers found in this study come specifically from this region [6,11].Currently, there are only seven Disa species from Zambia and surrounding countries registered on theglobal IUCN Red List and no species from other genera used for chikanda [54]. Most of the orchidspecies used for chikanda seem to have a widespread distribution, but local populations as well asendemic species could be at risk of overharvesting, and we urge the addition of the most frequentlytraded chikanda species, such as Disa robusta and Satyrium buchananii, to the IUCN Red List. Althoughno reduction in commerce is evident, the people involved in the chikanda trade seem genuinelyconcerned about the welfare of local orchids and to be interested in exploring other options. Sincethe chikanda harvesters especially seem to be in a vulnerable position, where they have to rely onsurrounding natural resources to secure their livelihoods [15], it is essential that, when trying to protectorchids used for chikanda, the situation of the people dependent on the trade be taken into account aswell. Currently, the development of sustainable cultivation of chikanda orchids is being attemptedin collaboration with the Cape Institute of Micropropagation (Barrydale, South Africa), and possiblealternative sources of income for the people involved in chikanda trade, such as honey production,are being explored [55]. Alternatively, since the purpose of the chikanda orchids is mainly to bindand create an elastic structure for the cake, it might be possible to encourage the use of an alternativesource of starch to replace the tubers.
5. Conclusions
DNA barcoding using the nuclear ribosomal ITS marker proved to be useful in identifyingterrestrial orchid species traded as chikanda on local Zambian markets and outperformed identificationusing the core land-plant barcoding markers matK and rbcL. Sixteen orchid species, including threepreviously undocumented ones, were identified from marketed chikanda tubers, bringing the totalnumber of orchid species used for chikanda to at least 49. The species most frequently found onthe markets were Disa robusta, Satyrium buchananii and Platycoryne crocea. However, the results areonly as good at the reference material, and an expanded reference database in combination withan underpinned phylogenetic framework for Habenaria and related genera would likely improvethe reliability of the identifications. Tubers are harvested from various regions in Zambia andTanzania, and additional international chikanda trade hubs have been identified on the border with theDemocratic Republic of Congo and Angola. People involved in the chikanda trade indicate that bothorchid quality and quantity are decreasing and are willing to consider alternatives to the chikandatrade to secure their income.
Supplementary Materials: The following supplementary figures and tables are available online at http://www.mdpi.com/2073-4425/9/12/595/s1, Figure S1. bPTP analysis of all chikanda matK query and reference samples;query samples identifiers are the collection number and the BLAST identification; reference samples from GenBankhave their 2-letter, 6-numeral accession numbers as identifier; other reference samples have their collection numberand identification based on morphology as identifier. Figure S2. bPTP analysis of all chikanda rbcL query andreference samples; query samples identifiers are the collection number and the BLAST identification; reference
Genes 2018, 9, 595 13 of 15
samples from GenBank have their 2-letter, 6-numeral accession numbers as identifier; other reference sampleshave their collection number and identification based on morphology as identifier. Figure S3. bPTP analysisof all chikanda nrITS query and reference samples; query samples identifiers are the collection number andthe BLAST identification; reference samples from GenBank have their 2-letter, 6-numeral accession numbersas identifier; other reference samples have their collection number and identification based on morphology asidentifier. Table S1. Brahms RDE file of novel voucher specimens of orchid taxa samples in this study; for athorough explanation of the column headers we refer to the online support material of the BRAHMS project [56].Table S2. Genbank accession numbers of the chikanda tuber sequences. Table S3. Hit table with the first 5 BLASTtop hits per sample for nrITS. Table S4. Hit table with the first 5 BLAST top hits per sample for matK. Table S5. Hittable with the first 5 BLAST top hits per sample for rbcL. Table S6. Successfully sequenced samples with theirvernacular name, reported origin, identification based on sequence-similarity for nrITS and matK, identificationbased on the tree-based bPTP analysis and the consensus ID. Table S7. Orchid species used for chikanda accordingto literature.
Author Contributions: S.V., R.B. and H.d.B. conceived and designed the study in collaboration with S.J.K., D.C.,R.V. and N.W.; S.J.K. performed the chikanda collection and interviews in collaboration with D.C., G.M., N.W.and R.V; R.B., N.W. and K.Y. collected reference vouchers in the field, which were identified in collaborationwith B.B.; B.B, G.N and F.M. provided the reference sequence database for Habenaria species and some speciesin closely related genera; S.J.K. performed the labwork for the chikanda collections and M.B.F. for the referencecollections; S.J.K. performed the data analysis under supervision of S.V. and H.d.B.; S.V. and S.J.K. wrote themanuscript in consultation with the other authors and the manuscript was edited and reviewed in detail by B.G.,T.v.A. and H.d.B.
Funding: This research was funded by the Darwin Initiative (UK Government) Grant No. 23034 and NetherlandsOrganisation for Scientific Research (NWO) TASENE grant W 02.29.102 and Swedish Research Council grantsVR-SRL D0664801 and VR-UF E0347601.
Acknowledgments: The authors would like to thank the participants in the study for sharing their knowledgeand Simon Hultby from the Uppsala Botanical Garden for growing the sprouted chikanda orchid tubers.
Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of thestudy; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision topublish the results.
References
1. Kasparek, M.; Grimm, U. European trade in Turkish Salep with special reference to Germany. Econ. Bot.1999, 53, 396–406. [CrossRef]
2. Ece Tamer, C.; Karaman, B.; Utku Copur, O. A traditional Turkish beverage: Salep. Food Rev. Int. 2006, 22,43–50. [CrossRef]
3. Kreziou, A.; de Boer, H.; Gravendeel, B. Harvesting of salep orchids in north-western Greece continues tothreaten natural populations. Oryx 2015, 1–4. [CrossRef]
4. Ghorbani, A.; Gravendeel, B.; Selliah, S.; Zarré, S.; Boer, H. DNA barcoding of tuberous Orchidoideae: aresource for identification of orchids used in Salep. Mol. Ecol. Resour. 2017, 17, 342–352. [CrossRef] [PubMed]
5. Davenport, T.R.B.; Ndangalasi, H.J. Orchid Harvest—An Assessment of the Harvesting and Trade of Orchid TubersAcross Tanzania’S Southern Highlands; Wildlife Conservation Society: New York, NY, USA, 2001; p. 23.
6. Bingham, M.G. Chikanda trade in Zambia. Orchid Conserv. News 2004, 4, 22–25.7. Veldman, S.; Otieno, J.N.; van Andel, T.; Gravendeel, B.; de Boer, H.J. Efforts urged to tackle thriving illegal
orchid trade in Tanzania and Zambia for chikanda production. TRAFFIC Bull. 2014, 26, 47–50.8. Richards, A.I. Land, Labour and Diet in Northern Rhodesia: An Economic Study of the Bemba Tribe; Oxford
University Press: Oxford, UK, 1939.9. Bingham, M.G. Chikanda an unsustainable industry. The Lowdown Magazine. 2007, 4.10. Temzie Bites Zambian Food|Chikanda African Ham (Polony) Recipe. 2017. Available online: https:
//www.youtube.com/watch?v=mDGKx-GOLoA (accessed on 24 October 2018).11. Davenport, T.R.B.; Ndangalasi, H.J. An escalating trade in orchid tubers across Tanzania’s Southern
Highlands: assessment, dynamics and conservation implications. Oryx 2003, 37, 55–61. [CrossRef]12. Cribb, P.J.; Leedal, G.P. The mountain Flowers of Southern Tanzania: A Field Guide to the Common Flowers;
AA Balkema: Rotterdam, NL, USA, 1982.13. Nyomora, A.M.S. Distribution and abundance of the edible orchids of the Southern Highlands of Tanzania.
Tanzanian J. Sci. 2005, 31, 45–54. [CrossRef]
Genes 2018, 9, 595 14 of 15
14. Mapunda, L.N.D. Edible Orchids in Makete District, the Southern Highlands of Tanzania: Distribution,Population and Status. Master’s Thesis, Swedish Biodiversity Centre, Uppsala Universitet (Sweden),Uppsala, Sweden, 2007.
15. Challe, J.F.; Price, L.L. Endangered edible orchids and vulnerable gatherers in the context of HIV/AIDS inthe Southern Highlands of Tanzania. J. Ethnobiol. Ethnomedicine 2009, 5, 41. [CrossRef] [PubMed]
16. Kasulo, V.; Mwabumba, L.; Munthali, C. A review of edible orchids in Malawi. J. Hortic. For. 2009, 1, 133–139.17. Hamisy, C.W. Development of conservation strategies for the wild edible orchid in Tanzania. Prog. Rep.
Rufford Small Grants Found. 2010.18. Veldman, S.; Gravendeel, B.; Otieno, J.N.; Lammers, Y.; Duijm, E.; Nieman, A.; Bytebier, B.; Ngugi, G.;
Martos, F.; van Andel, T.R.; et al. High-throughput sequencing of African chikanda cake highlightsconservation challenges in orchids. Biodivers. Conserv. 2017, 1–18. [CrossRef]
19. Hebert, P.D.N.; Cywinska, A.; Ball, S.; de Waard, J. Biological identifications through DNA barcodes. Proc. R.Soc. B 2003, 270, 313–322. [CrossRef] [PubMed]
20. Raclariu, A.C.; Heinrich, M.; Ichim, M.C.; Boer, H. Benefits and limitations of DNA barcoding andmetabarcoding in herbal product authentication. Phytochem. Anal. 2017. [CrossRef] [PubMed]
21. Martos, F.; Le Péchon, T.; Ngugi, G.; Pailler, T.; Bellstedt, D.U.; Linder, H.P.; Bytebier, B. Phylogeneticrelationships amongst the African genera of the tribe Orchideae (Orchidaceae). Manuscript in prep. 2018.
22. Goodman, L.A. Snowball sampling. Ann. Math. Stat. 1961, 148–170. [CrossRef]23. Dunning, L.T.; Savolainen, V. Broad-scale amplification of matK for DNA barcoding plants, a technical note:
AMPLIFICATION OF matK FOR DNA BARCODING PLANTS. Bot. J. Linn. Soc. 2010, 164, 1–9. [CrossRef]24. Shepherd, L.D.; McLay, T.G. Two micro-scale protocols for the isolation of DNA from polysaccharide-rich
plant tissue. J. Plant Res. 2011, 124, 311–314. [CrossRef] [PubMed]25. Kress, W.J.; Wurdack, K.J.; Zimmer, E.A.; Weigt, L.A.; Janzen, D.H. Use of DNA barcodes to identify flowering
plants. Proc. Natl. Acad. Sci. USA 2005, 102, 8369–8374. [CrossRef] [PubMed]26. Sun, Y.; Skinner, D.Z.; Liang, G.H.; Hulbert, S.H. Phylogenetic analysis of Sorghum and related taxa using
internal transcribed spacers of nuclear ribosomal DNA. Theor. Appl. Genet. 1994, 89. [CrossRef] [PubMed]27. Bonfield, J.K.; Smith, K.F.; Staden, R. A new DNA sequence assembly program. Nucleic Acids Res. 1995, 23,
4992–4999. [CrossRef] [PubMed]28. Staden, R. The Staden sequence analysis package. Mol. Biotechnol. 1996, 5, 233–241. [CrossRef] [PubMed]29. Larsson, A. AliView: A fast and lightweight alignment viewer and editor for large datasets. Bioinformatics
2014, 30, 3276–3278. [CrossRef] [PubMed]30. Zhang, J.; Kapli, P.; Pavlidis, P.; Stamatakis, A. A general species delimitation method with applications to
phylogenetic placements. Bioinformatics 2013, 29, 2869–2876. [CrossRef] [PubMed]31. Stamatakis, A.; Hoover, P.; Rougemont, J. A rapid bootstrap algorithm for the RAxML web servers. Syst. Biol.
2008, 57, 758–771. [CrossRef] [PubMed]32. Davenport, T.R.B.; Bytebier, B. Kitulo Plateau, Tanzania-a first African park for orchids. Orchid Rev. 2004,
112, 161–165.33. Boessenkool, S.; Mcglynn, G.; Epp, L.S.; Taylor, D.; Pimentel, M.; Gizaw, A.; Nemomissa, S.; Brochmann, C.;
Popp, M. Use of Ancient Sedimentary DNA as a Novel Conservation Tool for High-Altitude TropicalBiodiversity. Conserv. Biol. 2014, 28, 446–455. [CrossRef] [PubMed]
34. Van Geel, B.; Aptroot, A.; Baittinger, C.; Birks, H.H.; Bull, I.D.; Cross, H.B.; Evershed, R.P.; Gravendeel, B.;Kompanje, E.J.; Kuperus, P. The ecological implications of a Yakutian mammoth’s last meal. Quat. Res. 2008,69, 361–376. [CrossRef]
35. Kress, W.J.; Erickson, D.L. A two-locus global DNA barcode for land plants: the coding rbcL genecomplements the non-coding trnH-psbA spacer region. PLOS ONE 2007, 2, e508. [CrossRef] [PubMed]
36. Fazekas, A.J.; Burgess, K.S.; Kesanakurti, P.R.; Graham, S.W.; Newmaster, S.G.; Husband, B.C.; Percy, D.M.;Hajibabaei, M.; Barrett, S.C.H. Multiple multilocus DNA barcodes from the plastid genome discriminateplant species equally well. PLOS ONE 2008, 3, e2802. [CrossRef] [PubMed]
37. Kool, A.; de Boer, H.J.; Krüger, Å.; Rydberg, A.; Abbad, A.; Björk, L.; Martin, G. Molecular Identification ofCommercialized Medicinal Plants in Southern Morocco. PLoS ONE 2012, 7, e39459. [CrossRef] [PubMed]
38. Elder, J.F.; Turner, B.J. Concerted Evolution of Repetitive DNA Sequences in Eukaryotes. Q. Rev. Biol. 1995,70, 297–320. [CrossRef] [PubMed]
Genes 2018, 9, 595 15 of 15
39. Ganley, A.R.D.; Kobayashi, T. Highly efficient concerted evolution in the ribosomal DNA repeats: TotalrDNA repeat variation revealed by whole-genome shotgun sequence data. Genome Res. 2007, 17, 184–191.[CrossRef] [PubMed]
40. Harpke, D.; Peterson, A. Non-concerted ITS evolution in Mammillaria (Cactaceae). Mol. Phylogenet. Evol.2006, 41, 579–593. [CrossRef] [PubMed]
41. Feliner, G.; Rosselló, J. Better the devil you know? Guidelines for insightful utilization of nrDNA ITS inspecies-level evolutionary studies in plants. Mol. Phylogenet. Evol. 2007, 44, 911–919. [CrossRef] [PubMed]
42. Zheng, X.; Cai, D.; Yao, L.; Teng, Y. Non-concerted ITS evolution, early origin and phylogenetic utility of ITSpseudogenes in Pyrus. Mol. Phylogenet. Evol. 2008, 48, 892–903. [CrossRef] [PubMed]
43. Xu, B.; Zeng, X.-M.; Gao, X.-F.; Jin, D.-P.; Zhang, L.-B. ITS non-concerted evolution and rampant hybridizationin the legume genus Lespedeza (Fabaceae). Sci. Rep. 2017, 7, 40057. [CrossRef] [PubMed]
44. Chen, S.; Yao, H.; Han, J.; Liu, C.; Song, J.; Shi, L.; Zhu, Y.; Ma, X.; Gao, T.; Pang, X.; et al. Validation ofthe ITS2 region as a novel DNA barcode for identifying medicinal plant species. PLoS ONE 2010, 5, 1–8.[CrossRef] [PubMed]
45. Mezzasalma, V.; Bruni, I.; Fontana, D.; Galimberti, A.; Magoni, C.; Labra, M. A DNA barcoding approachfor identifying species in Amazonian traditional medicine: The case of Piri-Piri. Plant Gene 2017, 9, 1–5.[CrossRef]
46. Bytebier, B.; Dirk, U.; Bellstedt, H. Peter Linder A New Phylogeny-Based Sectional Classification for theLarge African Orchid Genus Disa. Taxon 2008, 57, 1233–1251.
47. Van der Niet, T.; Linder, H.P. Dealing with incongruence in the quest for the species tree: A case study fromthe orchid genus Satyrium. Mol. Phylogenet. Evol. 2008, 47, 154–174. [CrossRef] [PubMed]
48. Whitten, W.M.; Blanco, M.A.; Williams, N.H.; Koehler, S.; Carnevali, G.; Singer, R.B.; Endara, L.; Neubig, K.M.Molecular phylogenetics of Maxillaria and related genera (Orchidaceae: Cymbidieae) based on combinedmolecular data sets. Am. J. Bot. 2007, 94, 1860–1889. [CrossRef] [PubMed]
49. Johnson, L.A.; Soltis, D.E. matK DNA Sequences and Phylogenetic Reconstruction in Saxifragaceae s. str.Syst. Bot. 1994, 19, 143. [CrossRef]
50. Pridgeon, A.M.; Solano, R.; Chase, M.W. Phylogenetic relationships in Pleurothallidinae (Orchidaceae):Combined evidence from nuclear and plastid DNA sequences. Am. J. Bot. 2001, 88, 2286–2308. [CrossRef][PubMed]
51. Cunningham, A.B. Applied Ethnobotany: People, Wild Plant Use and Conservation; Earthscan: London, UK, 2001;ISBN 1853836974.
52. Martin, G.J. Ethnobotany: A Methods Manual; Earthscan: London, UK, 2004.53. Bingham, M.G.; Smith, P.P. Southern African Plant Red Data Lists: Zambia; Golding, J.S., Bandeira, S.O., Eds.;
SABONET: Pretoria, South Africa, 2002.54. IUCN IUCN Red List of Threatened Species. Version 2017-3. 2018.55. Bone, R.E.; Wightman, N.; Vinya, R.; Veldman, S.; Yokoya, K.; Hargreaves, S.; Kendon, J.; Crous, H.
Edible wild orchid trade: sustaining livelihoods and biodiversity in Zambia. Darwin Initiative MainProject Annual Report. 2017.
56. Filer, D. BRAHMS V5. 6202 Botanical Research and Herbarium Management System (Software). 2007.
© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open accessarticle distributed under the terms and conditions of the Creative Commons Attribution(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
